PROJECT SUMMARY/ABSTRACT Claiming more than 40,000 lives in the United States in 2015, breast cancer presents an important health focus. Mammography and ultrasound, current screening methods, suffer from low sensitivity and low positive predictive value, respectively, particularly in patients with dense breast tissues. Therefore, a non-invasive method of distinguishing between benign and malignant lesions that could be incorporated with current screening modalities is critically needed. With more advanced screening methods, there is an increase in the detection of early malignant lesions, for which breast-conserving treatment has become more routine. However, intraoperative frozen-section margin assessment is time consuming and suffers from low sensitivity, while post-operative histological analysis leaves potential for positive margins, strongly correlated with reoccurrence. Therefore, a real-time method to detect tumor margins intraoperatively is critically needed. We propose using spectroscopic photoacoustic and fluorescence molecular imaging combined with a clinically- translatable contrast agent targeted to a novel breast cancer marker (B7-H3) to non-invasively distinguish normal from malignant tissues both during screening (aim 1) and intraoperatively during surgical resection (aim 3). The sensitivity of this imaging method will be increased with the use of machine learning post-processing algorithms to autonomously detect molecular B7-H3 signal (aim 2). In summary, this proposal will result in significant change to current clinical breast imaging and surgical resection practice to improve the detection and treatment of focal breast lesions. The training portion of this plan, required to accomplish these research goals, has been designed with trainee mentors with specific technical expertise. Dr. Willmann is an expert in translational molecular imaging and contrast agent use, while Dr. Rubin is an expert in bioinformatics, image processing, and machine learning for medical imagine purposes. Additionally, the project is supported by a diverse advisory committee with experts in clinical breast imaging (Dr. Debra Ikeda), optical imaging and intraoperative guidance (Dr. Christopher Contag), and clinical breast surgery (Dr. Irene Wapnir). To date, the candidate has developed expertise in photoacoustic, ultrasound, and fluorescence molecular imaging and molecular contrast agent development and in vivo use during her graduate and postdoctoral research. Her long term career goals include developing clinically translatable combined spectroscopic photoacoustic and fluorescence molecular imaging methods combined with novel contrast agents for cancer detection and differentiation. Additionally, her research will focus on developing machine learning algorithms for increasing the sensitivity of the molecular imaging approach as well as adapting her method for therapeutic purposes. In preparation for her independent research career, the training plan includes formal education in machine learning, digital signal processing, optical imaging, and cancer biology, as well as in career development classes and ethical conduct of research.